Special Guest Lecturer

Nonlinear Filtering and Sequential Analysis for Hidden Markov ModelsProf. B.L. Rozovsky
Professor of Mathematics and Aerospace and Mechanical Engineering Center for Applied Mathematical SciencesUSC I will discuss some new statistical algorithms that proved to be efficient in a range of emerging applications. These include: computer vision, target tracking in a video-stream, intrusion detection in high speed computer networks, etc.
The first part of the talk is concerned with recent progress in theory and applications of nonlinear filtering. I will argue that banks of interacting (nonlinear) Bayesian matched filters can be successfully utilized to handle dim and acutely maneuvering targets.
The potential range of applications of the developed spatial-temporal tracking algorithms includes a new generation of tracking systems for missile defense, terrestrial reconnaissance, machine vision systems, and certain medical imaging applications.
In the second part of the talk I will discuss an efficient anomaly detection system that detects denial-of-service attacks at multi-gigabit speeds with minimal detection delay for a given low false alarm rate. The proposed constant false alarm rate detector utilizes ideas of sequential statistical analysis. The algorithm is computationally simple and self-learning, it adapts well to various network loads and usage patterns. Biography:Dr. Boris Rozovsky is Professor of Mathematics and Aerospace and Mechanical Engineering at the University of Southern California. He is also Director of the USC Center for Applied Mathematical Sciences
His research interests are in the areas of stochastic systems, nonlinear filtering,
prediction and smoothing, inverse problems for randomly perturbed systems,
target tracking, intrusion detection and mathematical modeling of the Internet. He is the
author of four books in the general area of stochastic systems and more than 100 research
articles. He is a Fellow of Institute of Mathematical Statistics and a recipient of other awards. Refreshments will be served.

Location: Seaver Science Library (SSL) - 150

Audiences: Everyone Is Invited

Contact: Nancy Levien